Bug 1651056 - s3 deployment section in the OCS operations guide needs a update
Summary: s3 deployment section in the OCS operations guide needs a update
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: Red Hat Gluster Storage
Classification: Red Hat Storage
Component: doc-Container_Native_Storage_with_OpenShift
Version: ocs-3.11
Hardware: Unspecified
OS: Unspecified
unspecified
high
Target Milestone: ---
: OCS 3.11.z Batch Update 2
Assignee: Chandrakanth Pai
QA Contact: Pulkit Kundra
URL:
Whiteboard:
Depends On:
Blocks: 1672844
TreeView+ depends on / blocked
 
Reported: 2018-11-19 05:01 UTC by Ashmitha Ambastha
Modified: 2019-04-01 06:08 UTC (History)
14 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2019-04-01 06:08:09 UTC
Embargoed:


Attachments (Terms of Use)
PDF version of the OCS 3.10 Operations Guide (7.30 MB, application/pdf)
2019-03-18 23:36 UTC, Chandrakanth Pai
no flags Details

Description Ashmitha Ambastha 2018-11-19 05:01:37 UTC
Description of problem:
The s3 deployment using the steps mentioned in the operations guide section 5.1 needs an update with respect to the commands and examples given.
Link to the guide :  https://access.redhat.com/documentation/en-us/red_hat_openshift_container_storage/3.11/html-single/operations_guide/#S3_Object_Store

1. In step 1, there has been no key given in the example. 
2. In step 1 again, the namespace name should be maintained the same with respect to the deployment guide. It's better to have uniformity. 
3. In step 2, the example has NAMESPACE,HEKETI_URL, STORAGE_CLASS variables not added. It's the same command given twice. 
4. In step 2 again, how do we know what should the storage class name be. Example is required to explain this. 
5. In step 3, again the command to be run and the example given are the exact same and there is an indentation mistake.
6. In step 4, the note says that the S3_ACCOUNT name, S3_USER name, and S3_PASSWORD. PVC and META_PVC are obtained from the previous step, but doesn't mention how one can figure that out.  
7. In step 5, verification of the s3 deployment should be written correctly. For verification if the s3 pod is up and running, the command is wrong. That should be # oc get pods  command to see status of the pod.
 
Version-Release number of selected component (if applicable): v3.11 and backports if required.

How reproducible: always.

Actual results: S3 deployment was not successful. 

Expected results: S3 deployment should be successful.

Comment 2 Saravanakumar 2019-03-13 10:36:24 UTC
1. In step 1, there has been no key given in the example. 

// Update it as:

# oc create secret generic heketi-storage-project-admin-secret
--from-literal=key=abcd  --type=kubernetes.io/glusterfs

2. In step 1 again, the namespace name should be maintained the same with respect to the deployment guide. It's better to have uniformity. 

storage-project is used a namespace name in other places as well. So, should not be an issue.

3. In step 2, the example has NAMESPACE,HEKETI_URL, STORAGE_CLASS variables not added. It's the same command given twice. 

// Before example, update it as :
# sed -e 's/${HEKETI_URL}/<HEKETI_URL>/g'  -e 's/${STORAGE_CLASS}/<STORAGE_CLASSNAME>/g' -e  's/${NAMESPACE}/<NAMESPACE_NAME/g'   /usr/share/heketi/templates/gluster-s3-storageclass.yaml | oc create -f -

// Leave the example as it is.


4. In step 2 again, how do we know what should the storage class name be. Example is required to explain this. 

Storage class here is new storage class being created. You are giving a new name to the storage class being created.
The step (Create a GlusterFS StorageClass file. Use the HEKETI_URL and NAMESPACE from the current setup and set a STORAGE_CLASS name.) is clear.


5. In step 3, again the command to be run and the example given are the exact same and there is an indentation mistake.

// Before example, update it as:
sed -e 's/${VOLUME_CAPACITY}/<NEW SIZE in Gi>/g'  -e  's/${STORAGE_CLASS}/<STORAGE_CLASSNAME>/g'  /usr/share/heketi/templates/gluster-s3-pvcs.yaml | oc create -f -

// Leave example as is.


6. In step 4, the note says that the S3_ACCOUNT name, S3_USER name, and S3_PASSWORD. PVC and META_PVC are obtained from the previous step, but doesn't mention how one can figure that out.  

// The note is clear
// (Set the S3_ACCOUNT name, S3_USER name, and S3_PASSWORD. PVC and META_PVC are obtained from the previous step.) 

// maybe better avoid it the "NOTE" and include the NOTE as part of documentation itself.

7. In step 5, verification of the s3 deployment should be written correctly. For verification if the s3 pod is up and running, the command is wrong. That should be # oc get pods  command to see status of the pod.

// Change the Step 5 output as:

# oc get pods -o wide
NAME                             READY     STATUS    RESTARTS   AGE       IP             NODE
gluster-s3-azkys                 1/1       Running   0          4m        10.130.0.29    node3
..

Comment 5 RamaKasturi 2019-03-15 13:48:26 UTC
Hello Anjana,

   Has this fix been made available @ 3.10 docs as well ? Asking because i do not see a 3.10 doc link being provided here

Thanks
kasturi

Comment 8 Chandrakanth Pai 2019-03-18 23:36:54 UTC
Created attachment 1545457 [details]
PDF version of the OCS 3.10 Operations Guide


Note You need to log in before you can comment on or make changes to this bug.